Objective: To evaluate whether rheumatoid arthritis (RA) is associated with a characteristic clinical pattern of pain which may be useful as a criterion to differentiate RA from other rheumatic diseases.
Methods: 2300 patients from the ReumaLink data bank project with definite rheumatic diseases were studied. Of these 907 patients (39.5%) fulfilled the ARA/ACR revised criteria for RA, while 1393 had rheumatic diseases other than RA. The following diagnostic attributes of pain were considered: localization, symmetry, continuity, modulation, relationship with time and with loads/movements, tenderness.
Results: After a descriptive analysis, some pain characteristics were selected individually and others were combined. Only 8 variables were considered for a predictive analysis. Univariate analysis showed that symmetric pain is the most potent discriminating item, with 82.2% sensitivity, 69.2% specificity, a 61% positive predictive value and a 83.3% negative predictive value. A higher probability of RA was present in patients with symmetric pain than in those with asymmetric pain (odds ratio = 7.8). A multivariate analysis performed on 1627 patients showed that a specific clinical pattern of pain (symmetrical pain, pain following joint pressure, mainly present at night or in the morning, continuous) could predict RA patients with a 68.9% likelihood. The lack of these symptoms excluded RA with 92% probability.
Conclusion: The clinical pattern of pain defined by us can predict RA with a 70% probability. This value reaches 86% when the variables "pain in a fixed joint" and "pain decreased by load/movements" are added. These results indicate that determining the clinical pattern of pain is a useful screening tool for suspected RA, in particular early in the disease course.